What a trading strategy makes per trade on average, once its wins and losses are netted out.
Expectancy is what your strategy makes on an average trade once you net the wins against the losses. In R it is written as EV per trade, so a strategy that averages +0.2R nets you a fifth of your risk every time you click the button. Positive expectancy is the whole game. A strategy can win 8 of its 10 trades and still bleed money if the 2 losses are bigger than the 8 wins, which is exactly why win rate fools so many people.
Expectancy is the average result of a trade with wins and losses combined. Multiply your win rate by your average win, subtract your loss rate times your average loss, and the number you are left with is what one trade is worth before you place it.
In R terms that average is your EV per trade, and it is the cleanest one number summary of whether an edge exists. A coin flip with a 1 to 1 payoff sits at zero. Anything reliably above zero, after costs, is an edge worth trading.
Expectancy says nothing about the ride, though. A system can carry a healthy +0.3R and still hand you a 15 trade losing streak that knocks you out of the seat before the math has time to pay you back.
Because win rate hides the size of your losses. You can win 70% of the time and still lose money if your average loss is three times your average win.
Expectancy folds both numbers into one. A 35% win rate trend system running at +0.6R will quietly beat a 70% win rate scalper at +0.05R, even though the scalper feels far better to trade. The market pays you on expectancy, not on how often you are right. For more on telling a real edge from a hot streak, read is your edge real or luck.
There is no universal number, because it scales with how often you trade. A +0.05R edge over 2,000 trades a year is a serious machine. The same +0.05R over 40 trades a year is noise you cannot tell apart from luck.
As a rough read, most tradeable retail systems land between +0.1R and +0.4R per trade after costs. Above +0.5R, check your data before you celebrate, because that is more often a sign of a tiny sample or survivorship than a genuine edge.
Quantprove reads your EV per trade straight from your closed trades and feeds it into your Edge Score as one of the core edge magnitude inputs, capped so a couple of monster trades cannot run the number away.
It also wants a real sample before it trusts the figure. Expectancy from 30 trades and expectancy from 500 trades are not the same evidence, so the score is scaled down until your sample earns full weight near 500 trades. Expectancy travels with your R multiple data, so the two are best read together.